The AI Security Panic — When AI Stops Advising and Starts Acting
Unbearably Bullish with Donald B. Johns III · Sep 16, 2026 · 35 min
Show Notes
Episode 11 of Wall Street Wednesday on Unbearably Bullish — when artificial intelligence stops advising and starts acting, the security implications change everything for investors, institutions, and markets.
- THE AI SECURITY PANIC:
Welcome to Wall Street Wednesday #11 on the Unbearably Bullish Podcast.
In today's rapidly evolving technological landscape, the integration of artificial intelligence into business operations has become a defining factor for success. However, the security reality of AI raises critical questions about whether this emerging technology is truly working for us, or if it poses unforeseen risks. Join Donald Johns from The Bull You Can't Bear, Unbearably Bullish Podcast, as he delves into the intricacies of artificial intelligence, its implications on business leadership, and the importance of understanding its role in finance and investing. This discussion is tailored for entrepreneurs, investors, and anyone interested in financial education, stock market analysis, and wealth building, providing insights into the investor mindset and success strategies in the face of tech trends. By exploring the potential benefits and pitfalls of AI, viewers will gain a deeper understanding of how to harness its power effectively, fostering a stronger, more secure approach to personal finance, market analysis, and business growth, all while maintaining an unbearably bullish outlook o
You’ll learn about
- Reported AI-agent security incidents involving OpenAI and Anthropic
- Chatbots versus autonomous AI agents
- Direct and indirect prompt-injection attacks
- Credentials, permissions and excessive agency
- AI-enabled cybercrime and cyberespionage
- Deepfakes, voice cloning and digital impersonation
- Model theft, distillation and capability extraction
- Enterprise AI security, governance and liability
- The potential cybersecurity investment opportunity
- Bull versus Bear: Can AI defenders remain ahead of AI attackers?
- The AI Security Dashboard for businesses, investors and consumers
The episode also examines the different security challenges surrounding:
- OpenAI and ChatGPT — agent capability, containment and cyber evaluations
- Anthropic and Claude — agent behavior, cyber misuse and unauthorized-access reports
- Google and Gemini — prompt injection across connected productivity ecosystems
- xAI and Grok — persistent agents, deepfakes and real-time information distribution
- Microsoft, Meta and global AI laboratories — enterprise access, open models and international competition
This is not a claim that consumer chatbots have become conscious or are freely attacking the internet. Many reported incidents involved specialized models, reduced safeguards, cybersecurity evaluations or third-party systems.
But the warning is real
A chatbot generates an answer.
An agent may send the email.
Run the code.
Open the file.
Use the credential.
Or take the action.
The next phase of the AI race may be decided by security, control, accountability and trust—not intelligence alone.
Advice is information. Access is power. Autonomy is leverage.
- Like the episode
- Comment with the AI security risk that concerns you most
About Wall Street Wednesday
Markets. Investing. Wealth Building.
Market trends, investing principles, economic insights, and long-term wealth-building strategies through an educational lens.
Episode disclosure
This video is provided solely for general educational and entertainment purposes. Nothing presented constitutes financial, investment, legal, cybersecurity, privacy, accounting or tax advice; an offer or solicitation; or an endorsement or recommendation of any company, model, product, security or investment strategy. References to OpenAI, ChatGPT, Anthropic, Claude, Google, Gemini, xAI, Grok, Microsoft, Meta and other organizations are educational examples. Reported incidents may involve specialized research models, evaluation environments, reduced safeguards or third-party systems and should not be interpreted as proof that ordinary consumer products independently engaged in the same conduct. Information is based on publicly available materials as of September 16, 2026, may remain under investigation and may be revised. AI systems and security risks vary by model, configuration, tools, permissions and deployment. Conduct independent research and consult appropriately qualified professionals before making consequential technology, cybersecurity or investment decisions. Investing involves risk, including possible loss of principal. Past performance does not guarantee future results.
